Skill Detection on SkillTrustBench
98.48Loose F1AI-Infra-Guard
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AI-Infra-GuardBase model=Claude Opus 4.62026.06 | 98.48 | 97.25 | 99.74 | 6.63 | |
| AI-Infra-GuardBase model=GLM 5.12026.06 | 98.36 | 97.01 | 99.74 | 7.23 | |
| AI-Infra-GuardBase model=Gemini 3.5 Flash2026.06 | 97.92 | 99.47 | 96.41 | 1.2 | |
| AI-Infra-GuardBase model=Kimi 2.62026.06 | 97.8 | 98.95 | 96.67 | 2.41 | |
| AI-Infra-GuardBase model=DeepSeek v4 Flash2026.06 | 97.4 | 98.68 | 96.15 | 3.01 | |
| AI-Infra-GuardBase model=Hy3 Preview2026.06 | 97.14 | 98.68 | 95.64 | 3.01 | |
| AI-Infra-GuardBase model=MiniMax M2.72026.06 | 96.23 | 97.63 | 94.87 | 5.42 | |
| AI-Infra-GuardBase model=Mimo v2.52026.06 | 95.98 | 97.11 | 94.87 | 6.63 | |
| AI-Infra-GuardBase model=GPT 5.52026.06 | 95.66 | 92.57 | 98.97 | 18.67 |